Subtle genetic modifications transformed an enteropathogen into a flea-borne pathogen

Yersinia pestis is the etiological agent of plague, one of the most devastating diseases in human history. Because of the identification of the plague bacillus, the understanding of its epidemiological cycle, and the advent of efficient antibiotic therapies, the death toll due to the plague has dramatically decreased over the last 100 y. However, this disease has never been eradicated, and because of its rodent reservoir, it will not be eradicated in the near future. On the contrary, human plague cases have been reported since the 1990s in countries where the disease was thought to be extinct for decades, and therefore, the plague is now categorized as a reemerging disease. Y. pestis has two characteristics that distinguish it from most other bacterial pathogens. One is its exceptional pathogenicity for humans, with a mortality rate of 40–70% in ∼1 wk for the bubonic form and close to 100% in around 3 d for pneumonic plague. The other characteristic of Y. pestis is its transmission from rodent to rodent and from rodent to human by an infectious fleabite. These two characteristics are probably linked (see below), and therefore deciphering the mechanisms that explain the peculiar mode of transmission of Y. pestis by fleas is a key step in the understanding of pathogen evolution and gain of virulence. In PNAS, Chouikha and Hinnebusch identify a fundamental event in the adaptation of the plague bacillus to its flea vector (1).

Recent Y. pestis population genetics analyses have shown that the plague bacillus emerged very recently (∼3,000 y ago) from the enteropathogen Yersinia pseudotuberculosis (2, 3). The two species are genetically almost identical, with more than 97% nucleotide identity for most of their genes (4). However, the two species have dramatically different epidemiological and pathological features. Y. pseudotuberculosis, like other pathogens of the Enterobacteriaceae family, is transmitted by the oral route and causes a usually moderate infection of the digestive tract. In contrast, Y. pestis is transmitted by fleas and may be considered as one of the most pathogenic bacteria for humans. This extreme pathogenicity and flea-borne transmission were acquired after Y. pestis diverged from its Y. pseudotuberculosis ancestor. Surprisingly, this emergence was characterized by limited gene acquisition and massive gene inactivation, with the loss of 13% of the functions of the ancestral Y. pseudotuberculosis (due to mutations, deletions, or insertions) (4). The necessity of fleas for transmission created a selective pressure that certainly had a strong impact on the pathogenicity of

Chouikha and Hinnebusch identify a fundamental event in the adaptation of the plague bacillus to its flea vector.

this newly emerged arthropod-borne bacterium. Because fleas are blood-sucking insects, the new flea-rodent-flea epidemiological cycle of the plague bacillus necessitates the presence of bacteria in the blood stream to allow an efficient transmission from one host to another one. Furthermore, the number of circulating Y. pestis bacteria has to be high (>10⁷ cells/mL of blood) to allow an efficient transmission by fleas (5). This means that the species Y. pestis would not have been able to maintain itself if it had not acquired the capacity to cause a highly severe and systemic infection. Because transmission by fleas has been crucial for the evolution of the bacillus, a key question is as follows: how did Y. pestis become a flea-borne pathogen? A means to answer this question is to ask the reverse question: why do fleas not transmit Y. pseudotuberculosis? The Hinnebusch laboratory used this approach and performed elegant works, culminating with the one presented here, to shed light on the genetic modifications that led to the transformation of an enteropathogen into a flea-borne pathogen (Table 1).

Table 1.

Gain and loss of functions that were necessary for an efficient transmission of Y. pestis by fleas

Blue, gene inactivation that occurred in Y. pestis to favor its transmission by fleas; red gene, acquisition that occurred in Y. pestis to favor its transmission by fleas.

A prerequisite for arthropod-borne transmission is to maintain the vectorial capacity of the infected insect vector. However, when Xenopsylla cheopis (the rat flea, a major vector of plague) are fed on Y. pseudotuberculosis-contaminated blood, they develop symptoms, and ∼40% of them die rapidly (6). In contrast a Y. pestis-infected blood meal does not cause any toxicity to the fleas. The capacity of the plague bacillus to colonize its vector without killing it was therefore crucial for its new lifestyle. Chouikha and Hinnebusch identify the molecular bases for the nontoxicity of Y. pestis for fleas (1). Using a comparative proteomic approach on subfractions of Y. pestis and Y. pseudotuberculosis, the authors identify the urease enzyme as the toxic component. It has been known by microbiologists that one of the few distinctive phenotypic traits between the two species was their urease activity. However, no relationship with flea-borne transmission had ever been suggested. The ure locus is composed of seven genes, which are well conserved between the two species. However, the last gene of this locus (ureD) in Y. pestis contains a point mutation that leads to a truncated protein and to the absence of urease activity (7). Chouikha and Hinnebusch demonstrate that this mutation is responsible for the nontoxicity of Y. pestis to fleas. Deletion of ureD abolished the lethaliity of Y. pseudotuberculosis for two different flea species (X. cheopis and the ground squirrel flea Oropsylla montana), whereas introduction of a functional ureD in Y. pestis caused a rapid death of most fed insects. Interestingly, the inactivating mutation, which is due to an additional G in a tract of six G in the ureD sequence (7), was already present in the phylogenetically most ancient strains of Y. pestis, indicating that this mutation arose early after Y. pestis divergence from Y. pseudotuberculosis (1). This work thus provides the explanation for the early genetic change that transformed Y. pseudotuberculosis into a potentially arthropod-borne pathogen.

The second step for an efficient transmission by fleas requires the survival and multiplication of the bacteria in the digestive tract of the insect vector. Y. pestis emergence was accompanied by the acquisition of two specific plasmids. One, designated pFra (or pMT1), encodes a phospholipase D that Hinnebusch's laboratory showed to protect Y. pestis from the toxicity of the products resulting from the digestion of the blood meal in the midgut of the flea (8). Y. pseudotuberculosis, which does not harbor this plasmid, cannot survive in the flea midgut. Introduction of the phospholipase D-encoding gene into Y. pseudotuberculosis endowed the bacteria with a capacity to multiply in the flea midgut similar to that of Y. pestis (9). Therefore, another genetic change that increased the efficiency of Y. pestis flea-borne transmission was the acquisition of the pFra plasmid.

Different flea species have varying vectorial capacities and modes of Y. pestis transmission. For instance, O. montana, which is the primary flea vector for plague in North America, transmits Y. pestis, by the so-called “early phase” mechanism (10). However, the rat flea X. cheopis, which plays a major epidemiological role in various plague foci worldwide, uses another mechanisms known as “blockage” (11). Y. pestis cells form an aggregate in the proventriculus (flea foregut) that prevents the blood from flowing into the midgut during feeding and favors the regurgitation of portions of the bacterial aggregates into the dermis of the mammalian host. A third step for an efficient transmission of Y. pestis by X. cheopis relies on its capacity to block fleas. This blockage results from the formation of an extracellular matrix (a biofilm), which is encoded by the chromosomal operon hsmHFRS (12). Y. pseudotuberculosis carries an hmsHFRS operon identical to that of Y. pestis, but is unable to block fleas, indicating that other genes involved in biofilm formation may differ between the two species. In various bacteria, the formation of a biofilm directly depends on the intracellular concentration of cyclic-di-GMP. The Hinnebusch laboratory has recently shown that the difference in the blockage capacity of Y. pestis and Y. pseudotuberculosis is due to mutations in three Y. pestis chromosomal genes that hamper c-di-GMP production. The product of one of these genes (rcs) inhibits c-di-GMP synthesis, and the other two (PDE2 and PDE3) degrade c-di-GMP (9). All three genes are functional in Y. pseudotuberculosis, whereas their inactivation in Y. pestis enhances c-di-GMP–mediated biofilm formation. These three mutations are present in all Y. pestis strains, including the closest descent from Y. pseudotuberculosis in the phylogenetic tree, indicating that they have been acquired early after the emergence of the plague bacillus (9).

Another major difference between the two species is that Y. pestis has adopted a narrow epidemiological cycle (rodent-flea-rodent), whereas Y. pseudotuberculosis infects numerous animal species and persists and multiplies in the environment. Of note, the genome of the entomopathogen Photorhabdus luminescens contains numerous Y. pestis/Y. pseudotuberculosis homologs. Overall, 2,107 genes in P. luminescens have orthologs in the Y. pestis/Y. pseudotuberculosis genomes, and 77% of these orthologous genes are syntenic (13). This suggests that the Y. pseudotuberculosis ancestor might also have been an entomopathogen. Further arguing for this hypothesis is the presence of at least 21 genes in Y. pseudotuberculosis that potentially encode insecticidal toxins, including 7 genes of the Tc family found in P. luminescens (6). These Tc proteins of Y. pseudotuberculosis are toxic for Manduca sexta neonates (14). Also of interest is the observation that several genes or gene clusters in P. luminescens are homologous to virulence genes of Y. pseudotuberculosis. These include the loci encoding invasins (inv and ail), a type III secretion system harbored by the Yersinia virulence plasmid pYV, and the yersiniabactin siderophore (13). Based on these observations, a possible scenario would be that the Y. pseudotuberculosis ancestor was an entomopathogen that progressively adapted to larger hosts and subverted some of the virulence genes used to kill insects to cause diseases in mammals. However, when one Y. pseudotuberculosis evolved to become Y. pestis, these genes were deleterious for the new mode of transmission of this bacterium by fleas and had to be inactivated. Chouikha and Hinnebusch nicely deciphered the successive gain and loss of functions that have allowed the emergence of the flea-borne plague bacillus.